org.apache.ctakes.dictionary.lookup2.consumer

Class SemanticCleanupTermConsumer

java.lang.Object

org.apache.ctakes.dictionary.lookup2.consumer.AbstractTermConsumer

org.apache.ctakes.dictionary.lookup2.consumer.SemanticCleanupTermConsumer

All Implemented Interfaces:

TermConsumer

public class SemanticCleanupTermConsumer
extends AbstractTermConsumer

Refine a collection of dictionary terms to only contain the most specific variations: "colon cancer" instead of "cancer", performed by span inclusion / complete containment, not overlap Also a start at wsd by trim of overlapping terms of conflicting but related semantic group. In this incarnation, any sign / symptom that is within a disease / disorder is assumed to be less specific than the disease disorder and is discarded. In addition, any s/s or d/d that has the same span as an anatomical site is discarded.

Since:

10/24/2014

Version:

%I%

Author:

SPF , chip-nlp

Constructor Summary

Constructors

Constructor and Description
SemanticCleanupTermConsumer(org.apache.uima.UimaContext uimaContext, Properties properties)

Method Summary

Modifier and Type	Method and Description
void	consumeHits(org.apache.uima.jcas.JCas jcas, RareWordDictionary dictionary, CollectionMap<TextSpan,Long,? extends Collection<Long>> textSpanCuis, CollectionMap<Long,Concept,? extends Collection<Concept>> cuiConcepts) Refine a collection of dictionary terms to only contain the most specific variations: "colon cancer" instead of "cancer", performed by span inclusion /complete containment, not overlap.
void	consumeTypeIdHits(org.apache.uima.jcas.JCas jcas, String defaultScheme, int cTakesSemantic, CollectionMap<TextSpan,Long,? extends Collection<Long>> semanticTerms, CollectionMap<Long,Concept,? extends Collection<Concept>> conceptMap)

Methods inherited from class org.apache.ctakes.dictionary.lookup2.consumer.AbstractTermConsumer

getCodingScheme, getUsedcTakesSemantics, hascTakesSemantic

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

SemanticCleanupTermConsumer

public SemanticCleanupTermConsumer(org.apache.uima.UimaContext uimaContext,
                                   Properties properties)

Method Detail

consumeHits

public void consumeHits(org.apache.uima.jcas.JCas jcas,
                        RareWordDictionary dictionary,
                        CollectionMap<TextSpan,Long,? extends Collection<Long>> textSpanCuis,
                        CollectionMap<Long,Concept,? extends Collection<Concept>> cuiConcepts)
                 throws org.apache.uima.analysis_engine.AnalysisEngineProcessException

Refine a collection of dictionary terms to only contain the most specific variations: "colon cancer" instead of "cancer", performed by span inclusion /complete containment, not overlap. For instance: "54 year old woman with left breast cancer." in the above sentence, "breast" as part of "breast cancer" is an anatomical site and should not be a S/S "Breast: "lump, cyst" in the above, breast is a list header, denoting findings on exam.

Specified by:

consumeHits in interface TermConsumer

Overrides:

consumeHits in class AbstractTermConsumer

Parameters:

jcas - -

dictionary - the dictionary: Anatomical Site, Disease/Disorder, Drug, combination, etc.

textSpanCuis - collection of discovered terms

cuiConcepts - map of cuis to concepts

Throws:

org.apache.uima.analysis_engine.AnalysisEngineProcessException

consumeTypeIdHits

public void consumeTypeIdHits(org.apache.uima.jcas.JCas jcas,
                              String defaultScheme,
                              int cTakesSemantic,
                              CollectionMap<TextSpan,Long,? extends Collection<Long>> semanticTerms,
                              CollectionMap<Long,Concept,? extends Collection<Concept>> conceptMap)
                       throws org.apache.uima.analysis_engine.AnalysisEngineProcessException

Parameters:

jcas - -

defaultScheme - - name of the coding scheme to use if no secondary (snomed, rxnorm) schemes exist

cTakesSemantic - cTakes IdentifiedAnnotation only accepts an integer as a cTakesSemantic

Throws:

org.apache.uima.analysis_engine.AnalysisEngineProcessException